Article Figures & Data
Figures
- FIG 1
Orientation and stoichiometry of the hu1B7 and putative hu11E6 epitopes. The hu1B7 antibody binds a single, well-defined epitope on the S1 subunit, while hu11E6 binds two highly homologous epitopes present on both the S2 and S3 subunits. (A) PTx can simultaneously engage antibodies binding the hu1B7 and hu11E6 epitopes. The data shown are from a sandwich ELISA in which one antibody was used to capture PTx (black symbols), the A subunit (S1-220K; gray symbols), or the B subunit (white symbols), which was then detected by a second antibody. Tested antibody pairs included an m1B7 capture antibody with a hu1B7 detection antibody (m1B7/hu1B7; circles), m1B7 capture with hu11E6 detection (m1B7/hu11E6; squares), and m11E6 capture with hu11E6 detection (m11E6/hu11E6; triangles). Error bars show the ranges for two duplicate samples in a single experiment. (B) PTx model showing the relative locations of the hu1B7 and hu11E6 epitopes. The experimentally defined hu1B7 epitope is shown in dark blue and the two predicted hu11E6 epitopes in light green. The lines extending from the epitopes are normal to the average plane, approximating the angle at which a bound Fab would project. A full-length huIgG1 structure (PDB entry 1HZH) (60) with the antigen binding sites highlighted in red is shown at the same scale. Graphics were generated with PyMol.
- FIG 2
Production of bispecific hu1B7/hu11E6 antibody. (A) Schematic overview of the process used to create the bispecific antibody. Single amino acid residue changes were introduced into the parent antibody Fc domains to create “knob” (T366Y; hu1B7H+) and “hole” (Y407T; hu11E6H−) variants, which were expressed and purified separately. An equimolar mixture of the two proteins was then subjected to a controlled reduction and reoxidation reaction, resulting in efficient formation of the heterodimeric, bispecific antibody. (B) The purity of the knob and hole variants and 2-MEA-catalyzed recombination were monitored by nonreducing SDS-PAGE. Samples shown included (+) or did not include (−) the three reaction components, i.e., hu1B7H+, hu11E6H−, and the 2-MEA redox step. (C) Purified hu1B7H+, hu11E6H−, and the bispecific antibody (bsAb) were digested with the IdeS enzyme to yield F(ab′)2 fragments. LC-MS was used to determine the purity of the bispecific preparation.
- FIG 3
Biophysical characterization of bispecific and parent antibodies. (A) Reducing and nonreducing SDS-PAGE analyses were used to assess the size and purity of the hu1B7 and hu11E6 full-length and Fab antibody fragments as well as the hu1B7/hu11E6 bispecific antibody. A total of 3 μg of protein was loaded per lane. (B) Differential scanning fluorimetry was performed to determine the melting profiles of full-length hu1B7 and hu11E6 (blue and green solid lines, respectively), the hu1B7 and hu11E6 Fabs (blue and green dashed lines, respectively), and the hu1B7/hu11E6 bispecific antibody (gray line). Samples were prepared in PBS, pH 7.4. Derivative data [d(fluorescence)/d(temperature)] are aligned below the raw melting curves to facilitate identification of transition temperatures. Curves show averages for three replicates at 400 μg/ml.
- FIG 4
Biochemical characterization of bispecific and parent antibodies. (A) The hu1B7/hu11E6 bispecific antibody (gray squares) was assessed for the presence of each unique binding site and the ability to simultaneously engage two different epitopes in a sandwich ELISA. The purified PTx-B subunit was used to capture the hu11E6 paratope, followed by biotinylated S1-220K and streptavidin-HRP to detect the hu1B7 paratope. The monospecific parent antibodies hu1B7H+ (solid circles) and hu11E6H− (hollow circles) were used as controls. (B) ELISA was used to confirm the PTx binding activity of all antibody variants and controls. A PTx-coated plate was incubated with dilutions of full-length hu1B7 or hu11E6 (solid circles with blue line and hollow circles with green line, respectively), the hu1B7 or hu11E6 Fab (solid triangles with dashed blue line and hollow triangles with dashed green line, respectively) or the hu1B7/hu11E6 bispecific antibody (squares with solid gray line), followed by anti-human kappa–HRP to detect bound chains. (C) A competition ELISA was used to determine the solution-based equilibrium dissociation constants (KD) for all variants. Data are shown for the full-length (solid bars) and Fab (striped bars) antibody formats. For all panels, the errors shown are standard deviations. The competition ELISA data were averaged for six replicates over three experiments. Statistical significance was determined using one-way ANOVA and Tukey's test. Statistical significance is indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001. In addition to the comparisons shown, the hu11E6 Fab fragment was significantly different from all other groups (****), and the hu1B7 Fab-hu11E6 Fab mixture was significantly different from the hu1B7 antibody (*).
- FIG 5
In vitro PTx neutralization measured by inhibition of CHO cell clustering. Adherent CHO-K1 cells were grown in the presence of 4 pM PTx preequilibrated with antibody dilutions. The lowest molar antibody-to-toxin ratio able to fully prevent clustering was recorded as the neutralizing ratio. Data are shown for neutralization with the full-length and bispecific antibodies (solid bars) and their Fab fragments (striped bars). The hu1B7 Fab was not able to fully neutralize the toxin at any concentration tested. Data are presented as geometric means for six replicates over three experiments, with error bars indicating 95% confidence intervals. Statistical analysis was performed using Kruskal-Wallace and Dunn's post hoc tests. Statistical significance is indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001.
- FIG 6
Antibody-mediated suppression of PTx-induced leukocytosis in vivo. (A) Antibody treatment alone does not raise WBC counts. Five-week-old female BALB/c mice were administered 200 μg hu1B7 or an equal volume of PBS via intraperitoneal injection, and the CD45+ WBC count 4 days after treatment was measured by flow cytometry. WBC counts of these and other mice were taken on day 0 to assess the baseline variability. (B) Antibody treatment suppresses PTx-induced leukocytosis. Mice were administered equal volumes of PBS, 2 μg PTx in PBS, or 2 μg PTx plus 20 μg total antibody (hu1B7, hu11E6, hu1B7 plus hu11E6, or the bispecific antibody) in PBS via a lateral tail vein injection, and CD45+ WBC counts were measured 4 days later by flow cytometry. Hollow symbols indicate the WBC counts recorded for baseline and the PBS- and PTx-only-treated control groups, and solid symbols indicate antibody-treated groups. Triangles represent groups that received PTx. Groups were compared using one-way ANOVA and Tukey's test, with statistically significant comparisons indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001. The comparison of the hu1B7-hu11E6 binary mixture and the hu1B7/hu11E6 bispecific antibody was not significant.
